The thesis focuses on the development of statistical theory in experimental design with applications in global optimization. It consists of four parts. In the first part, a criterion of design efficiency, under model uncertainty, is studied with reference to possibly nonregular fractions of general factorials. The results are followed by a numerical study and the findings are compared with those based on other design criteria. In the second part, optimal designs are dentified using Bayesian methods. This work is linked with response surface methodology where the first step is to perform factor screening, followed by response surface exploration using different experiment plans. A Bayesian analysis approach is used that aims to achieve both goals using one experiment design. In addition we use a Bayesian design criterion, based on the priors for the analysis approach. This creates an integrated design and analysis framework. To distinguish between competing models, the HD criterion is used, which is based on the pairwise Hellinger distance between predictive densities. Mixed-level fractional factorial designs are commonly used in practice but its aliasing relations have not been studied in full rigor. These designs take the form of a product array. Aliasing patterns of mixed level factorial designs are discussed in the third part. In the fourth part, design of experiment ideas are used to introduce a new global optimization technique called SELC (Sequential Elimination of Level Combinations), which is motivated by genetic algorithms but finds the optimum faster. The two key features of the SELC algorithm, namely, forbidden array and weighted mutation, enhance the performance of the search procedure. Illustration is given with the optimization of three functions, one of which is from Shekel's family. A real example on compound optimization is also given.

Modern reservoir management typically involves simulations of geological models to predict future recovery estimates, providing the economic assessment of different field development strategies. Integrating reservoir data is a vital step in developing reliable reservoir performance models. Currently, most effective strategies for traditional manual history matching commonly follow a structured approach with a sequence of adjustments from global to regional parameters, followed by local changes in model properties. In contrast, many of the recent automatic history matching methods utilize parameter sensitivities or gradients to directly update the fine-scale reservoir properties, often ignoring geological inconsistency. Therefore, there is need for combining elements of all of these scales in a seamless manner. We present a hierarchical streamline-assisted history matching, with a framework of global-local updates. A probabilistic approach, consisting of design of experiments, response surface methodology and the genetic algorithm, is used to understand the uncertainty in the large-scale static and dynamic parameters. This global update step is followed by a streamline-based model calibration for high resolution reservoir heterogeneity. This local update step assimilates dynamic production data. We apply the genetic global calibration to unconventional shale gas reservoir specifically we include stimulated reservoir volume as a constraint term in the data integration to improve history matching and reduce prediction uncertainty. We introduce a novel approach for efficiently computing well drainage volumes for shale gas wells with multistage fractures and fracture clusters, and we will filter stochastic shale gas reservoir models by comparing the computed drainage volume with the measured SRV within specified confidence limits. Finally, we demonstrate the value of integrating downhole temperature measurements as coarse-scale constraint during streamline-based history matching of dynamic production data. We first derive coarse-scale permeability trends in the reservoir from temperature data. The coarse information are then downscaled into fine scale permeability by sequential Gaussian simulation with block kriging, and updated by local-scale streamline-based history matching. he power and utility of our approaches have been demonstrated using both synthetic and field examples.

The first aim of this study is the development of extracellular recombinant therapeutic protease streptokinase producing Bacillus sp., and the second aim is to determine fermentation characteristics for streptokinase production. In this context, the signal (pre-) DNA sequence of B.licheniformis (DSM1969) extracellular serine alkaline protease enzyme gene (subC: Acc. No. X03341) was ligated to 5&rsquo / end of the streptokinase gene (skc: Acc. No. S46536) by SOE (Gene Splicing by Overlap Extension) method through PCR. The resulting hybrid gene pre(subC)::skc was cloned into the pUC19 plasmid. Then, the hybrid gene was sub-cloned to pMK4 plasmid which is an E. coli-Bacillus shuttle vector with high copy number and high stability. Recombinant plasmid pMK4::pre(subC)::skc was finally transferred into B. subtilis (npr- apr-) and B. licheniformis 749/C (ATCC 25972) species. Streptokinase production capacities of these two recombinant Bacillus species were compared. The highest production was observed in recombinant B. lichenifomis 749/C (ATCC 25972) strain in a defined medium which was optimized in terms of carbon and nitrogen sources by a statistical approach, namely Response Surface Methodology (RSM). RSM evaluated the streptokinase concentration as the response and the medium components as the independent variables. The highest recombinant streptokinase concentration was found as 0.0237 kgm-3 at glucose and (NH4)2HPO4 concentrations of 4.530 and 4.838 kgm-3 respectively. The fermentation and oxygen transfer characteristics of the streptokinase production were investigated in a 3 dm3 pilot scale batch bioreactor (Braun CT2-2) equipped with temperature, pH, foam, air inlet and agitation rate controls having a working volume of VR=1.65 dm3 using the production medium optimized for the recombinant B. lichenifomis 749/C (ATCC 25972) strain. Streptokinase and &amp / #946 / -lactamase activities, cell, glucose and organic acid concentrations, dissolved oxygen, pH, oxygen uptake rate, overall liquid phase mass transfer coefficient for oxygen, maintenance coefficient for oxygen, specific cell growth rate and yield coefficients were determined through the bioprocess. The bioprocess of recombinant streptokinase production was performed at uncontrolled pH of these bioreactor operation conditions: air inlet rate of Q0/VR=0.5 vvm, and the agitation rate of N=400min-1. The resulting streptokinase volumetric activity reached its maximum as 1.16 PUml-1 (0.0026 g/l streptokinase) at t=20 h.

Aspergillus fumigatus produces enzymes required for the hydrolysis of galactomannans like locust bean gum. Among these enzymes endo-beta-1,4 mannanase is also produced at high levels. However, the fungus is not safe for use in the food industry. Therefore, the gene encoding endo-beta-1,4-mannanase of A. fumigatus IMI 385708 was previously cloned in our laboratory into Aspergillus sojae ATCC11906 which is a safe microorganism for use in food applications. Altogether eight transformants were obtained. It was shown that some of these transformants overproduce the enzyme because of expression under the control of glyceraldehyde-3-phosphate dehydrogenase promoter and fusion to the glucoamylase signal and pro-peptide coding region of Aspergillus niger. In this study, mannanase production of these transformants was compared with A. fumigatus and A. sojae transformant AsT1 showed c. 12 fold increase with the maximum activity of 352 U/ml. The effects of initial medium pH and number of spores on activity were investigated and maximum activity was achieved at pH 7.0 and the number of spores was found as 3.6 &times / 106. Optimization of the growth conditions for maximum mannanase production in shake flasks by using the best mannanase producing transformant AsT1 was carried out by using Box-Behnken design under Response Surface Methodology. The highest beta-mannanase activity on the fourth day of cultivation at 30 &ordm / C was obtained as 363 U/ml in the optimized medium containing 7% sugar beet molasses, 0.43% NH4NO3, 0.1% K2HPO4, 0.05% MgSO4 as the weight/volume percentage at 207 rpm. On sixth day of cultivation under the optimized conditions, the highest mannanase activity was achieved as 482 U/ml which is 1.4 fold of 352 U/ml activity found on glucose medium previously. After 48 h of LBG hydrolysis by 40 U of mannanase, mannotriose, 61-galactosyl-beta-D-mannotriose and 63,64-di-alpha-galactosyl-beta-1,4-mannopentaose were found as the main products via HPLC analysis.

Kitchen waste, which is collected in large amounts from cafeterias, restaurants, dining halls, food processing plants, and household kitchens, have become a valuable material for bioprocess engineering. Due to the high carbohydrate fraction, kitchen waste has great potential to be used as a potential substrate for ethanol production. Utilization of it as a raw material in ethanol fermentation would also contribute to reduction of costs. In the first part of this study, the effect of pretreatment method and enzymatic hydrolysis on glucose production was evaluated. Dry baker&rsquo / s yeast, Saccharomyces cerevisiae, was used in fermentation experiments conducted with and without fermentation medium at pH 4.5 and 30oC for 48 hours. Close values of glucose concentration were obtained from no pretreated and hot water treated samples. The fermentation results indicated that ethanol can be produced at similar concentrations in bioreactors with and without fermentation medium addition (p &gt / 0.05). Thus, it is concluded that use of kitchen wastes as is disposed and without fermentation medium in ethanol fermentation could lower the cost to a large extent. In the second part of this study, the effects of solid load, which is proportional to the glucose concentration (10% to 20% (w/w)), inoculum level of Saccharomyces cerevisiae (5% to 15% (v/v)), and fermentation time (48 to 96 h) on production of bioethanol from kitchen waste were studied using Response Surface Methodology (RSM). A three-factor Box Behnken design was used. Ethanol concentration was used as a response in the resulting experimental design. High Pressure Liquid Chromatography (HPLC) method was used to determine ethanol and glucose concentrations. The statistical analysis of the constructed model developed by RSM suggested that linear effects of solid load, inoculum level, and fermentation time and quadratic effects of inoculum level and fermentation time were all significant (p &lt / 0.05) on bioethanol production. The model was verified by additional runs, which were not present in the design matrix. It was found that the constructed model could be used to determine successfully the bioethanol concentration with &gt / 90% precision. An optimum ethanol concentration of 32.16 g/L was suggested by the model with 20% (w/w) solid load, 8.85% (v/v) inoculum level and 58.8 hours of fermentation. Further study is needed to evaluate the optimal fermentation conditions in a large scale fermentation

In this thesis, several Reliability-based Design Optimization (RBDO) methods and algorithms for airplane floor grid layout optimization are proposed. A general RBDO process is proposed and validated by an example. Copula as a mathematical method to model random variable correlations is introduced to discover the correlations between random variables and to be applied in producing correlated data samples for Monte Carlo simulations. Based on Hasofer-Lind (HL) method, a correlated HL method is proposed to evaluate a reliability index under correlation. As an alternative method for computing a reliability index, the reliability index is interpreted as an optimization problem and two nonlinear programming algorithms are introduced to evaluate reliability index. To evaluate the reliability index by Monte Carlo simulation in a time efficient way, a kriging-based surrogate model is proposed and compared to the original model in terms of computing time. Since in RBDO optimization models the reliability constraint obtained by MCS does not have an analytical form, a kriging-based response surface is built. Kriging-based response surface models are usually segment functions that do not have a uniform expression over the design space; however, most optimization algorithms require a uniform expression for constraints. To solve this problem, a heuristic gradient-based direct searching algorithm is proposed. These methods and algorithms, together with the RBDO general process, are applied to the layout optimization of aircraft floor grid structural design.

現行不動產大量估價主要以特徵價格模型為基礎進行價格之預估，而常以鄰里、轄區或次市場虛擬變數或是與特定公共設施之距離作為控制區位價值之變數。惟僅以次市場變數之係數或是距離特定公共設施距離之係數衡量樣本之區位價值，則因係數之僵化性弱化或低估區位對不動產價格之影響，導致大量估價模型之精度難以突破。 本研究以區位價值波面之概念建立條件式殘差擬合變數，從空間角度評估各樣本之區位價值並以量化數值呈現各樣本區位價值之高低，在細膩處理區位價值下模型之預估能力相對提升。實證結果顯示，整體模型之絕對誤差平均值為10.1%，而10%、20%誤差命中率達62.9%、87.9%，相對優於過去研究之模型預估能力；另外，經過區域侷限性測驗發現，條件式殘差擬合變數修正模型不受次市場之侷限，對於是否劃分模型次市場已不影響模型之預估能力，且經由實證發現，當實價登錄樣本愈趨豐富時，模型之預估能力將更加提升，值得作為後續建立大量估價模型之參考。 / Hedonic model is the most commonly-used tool for real estate mass appraisal, and neighborhoods, districts or sub-market dummies or the distance from the specific public facilities are the common variables used to control the value of location in the model. However, controlling the location value by these ways leads to the coefficient rigidities, making it possible to underestimate the value of the location. This research sets up the conditional-selected residual fitting variable by the concept of location value response surface, and estimates the location value from the spatial perspective. The result shows that the MAPE of the model is 10.1%, and the hit-rate of 10% and 20% come to 62.9% and 87.9%, having significant improvement compared with the past studies. Besides, by the confinement test of sub-market, it has been proved that the CRF modified model successfully gets rid of confinement from the sub-market, and whether dividing sub-markets or not no longer affects the prediction capability of the model. Another test giving us new images that, when the train data gets richer as time goes, the prediction capability of the model gets higher as well.

The purpose of this study was to investigate perceptions of academic and social competence among adolescents with a continuum of inattentive and hyperactive/impulsive symptoms. Past literature suggests that children with Attention-Deficit/Hyperactivity Disorder (ADHD) display self-perceptions that are overly positive compared to external indicators of competence, a phenomenon that is referred to as the positive illusory bias (PIB; Owens, Goldfine, Evangelista, Hoza, & Kaiser, 2007). The PIB is well supported among children with ADHD, and recent research suggests that the PIB persists into adolescence. To date, research on the PIB has relied on difference scores (i.e., an indicator of competence is subtracted from student self-ratings); however, difference scores suffer from numerous methodological limitations (Edwards, 2002). The current study investigated the relationship between self and teacher ratings of academic and social competence and inattentive, hyperactive/impulsive, and overall ADHD symptoms among a diverse sample of 395 students and their teachers. Polynomial regression and response surface methods were used to account for self and teacher ratings separately and decrease reliance on differences scores. These methods have been recommended to answer complex questions related to agreement and disagreement between ratings. The results of this study suggest that some adolescents with ADHD symptoms demonstrate the PIB, while others perceive their impairments and rate themselves as having low competence aligned with teacher ratings. Accurate ratings of low competence were more common within the academic domain than the social domain for students with overall ADHD symptoms as well as specific inattentive and hyperactive/impulsive symptoms. Results within the social domain indicate that all ADHD symptoms increased more sharply as the discrepancy between self and teacher ratings increased. Student overestimation of competence in both the academic and social domains was shown to be more predictive of high inattentive symptoms compared to hyperactive/impulsive symptoms. These findings suggest this new analysis approach allowed for a more nuanced understanding of the complex relationship between student and teacher competence ratings and ADHD symptoms. Gaining a better understanding of the PIB through this improved methodology has the potential to influence assessment and intervention practices among school psychologists, and to contribute to future research in this area. This study contributes to the literature by being the first to (1) examine the PIB in relation to a range of general and specific ADHD symptoms, (2) use polynomial regression/response surface methods to address limitations of difference scores, and (3) explore the PIB among a school-based sample of adolescents.

Block copolymers (BCPs) are an attractive alternative for patterning applications used to produce next-generation microelectronic devices. Advancements require the development of high interaction parameter χ BCPs that enable patterning at the sub-10 nm length scale. Several organosilicon BCPs were designed to both enhance χ and impart an inherent etch selectivity that facilitates pattern transfer processes. Increasing the BCP silicon content both increases χ and bolsters the etch resistance, providing a pathway to designing new high-χ materials. Unfortunately, the BCPs investigated are not amenable to thermal annealing because the organosilicon block preferentially segregates to an air/vacuum interface and drives orientation parallel to the surface. A series of spin-coatable, polarity-switching top coats (as well as other strategies) were developed to provide a “neutral” top interface and promote the perpendicular orientation of BCP domains. In addition, a methodology for evaluating the neutral condition, relying on thickness quantization and the corresponding wetting behavior (i.e. island/hole topography) of lamellae. The top coat strategy was demonstrated for several BCP systems, and perpendicular structures can successfully be etched on commercial tools and be transferred into underlying substrates. The interaction parameter χ was evaluated using two methods to compare the performance of several BCPs: the order-disorder transition (ODT) of symmetric diblock copolymers, and the absolute scattering profile of a disordered BCP melt. Both methods, while severely limited for quantitative comparison, indicate trends towards higher χ with additional appended polar and organosilicon functional groups. Furthermore, the pattern fidelity is shown to be a function of the overall BCP segregation strength. The free energy of confined lamella was modeled algebraically to produce response surface plots capable of identifying process conditions favorable for perpendicular orientation. Thickness independent perpendicular orientation is only favorable using two neutral interfaces. Incommensurate film thicknesses are the most favorable, with commensurability conditions dependent on the wetting behavior at each interface. The modeling was supplemented with an extensive body of thin film experimental work that qualitatively agrees well with the above conclusions.

Unconfined multiple interacting confluent round jets are interesting from a purely scientific point of view, as interaction between neighboring jets brings additional complexity to the flow field. Unconfined confluent round jets also exist in various engineering applications, such as ventilation supply devices, sewage disposal systems, combustion burners, chemical mixing or chimney stacks. Even so, little scientific attention has been paid to unconfined confluent round jets. The present work uses both advanced measurement techniques and computational models to provide deeper understanding of the turbulent flow field development of unconfined confluent round jets. Both Laser Doppler Anemometry (LDA) and Particle Image Velocimetry (PIV) have been used to measure mean velocity and turbulence properties within two setups, consisting of a single row of 1×6 jets and a square array of 6×6 confluent jets. Simulations using computational fluid dynamics (CFD) of the 6×6 setup were conducted using three different Reynolds Averaged Navier-Stokes (RANS) turbulence models: the standard k-ε, the RNG k-ε and the Reynolds Stress model (RSM). The results from the CFD simulations were compared with experimental data. The employed RANS turbulence models were all capable of accurately predicting mean velocities and turbulent properties in the investigated confluent jet array. In general the RSM and k-ε std. models provided smaller deviations between numerical and experimental results than the RNG k-ε model. In terms of mean velocity the second-order closure model (RSM) was not found to be superior to the less complex standard k-ε model. The validated CFD model was employed in a parametrical investigation, including five independent variables: inlet velocity, nozzle diameter, nozzle edge-to-edge spacing, nozzle height and the number of jets in the array. The parametrical investigations made use of statistical methods in the form of response surface methodology. The derived response surface models provided information on the principal influence and relative importance of the investigated parameters within the investigated design space. The positions of the jets within the array strongly influence both mean velocity and turbulence. In all investigated setups the jets experience merging and combining. Square arrays also include considerable jet convergence, which was not present in the 1×6 jet array. Due to the jet convergence in square arrays the turbulent flow field, especially for jets far away from the array center, is affected by mean flow curvature. Jets located along the sides of square jet arrays experience strong jet-to-jet interactions that result in considerable jet deformation, shorter potential core, higher turbulent kinetic energy and faster velocity decay compared to other jets. Jets located at the corners of the array do not interact as strongly with neighboring jets as do the jets along the sides. The locations of merging and combined points differ considerably between different jets and different jet configurations. As the jets combine a zone with uniform stream-wise velocity and low turbulence intensity forms in the center of square jet arrays. This zone has been called Confluent Core Zone (CCZ) due to its similarities with the potential core zone of a single jet. Within the CCZ the appropriate scaling length changes from nozzle diameter to the effective source diameter. The parametrical investigation showed that nozzle diameter and edge-to-edge nozzle spacing were the most important of the investigated parameters, reflecting a strong dependence on dimensionless jet spacing, S/d0. Higher S/d0 delays both merging and combining of the jets and leads to a CCZ with lower velocity and longer downstream extension. Increasing the array size leads to a reduced combined point distance, a stronger inwards displacement of jets in the outer part of the array, and reduced entrainment near the nozzles. A higher inlet velocity was found to increase the jet convergence in the investigated square confluent jet arrays. Nozzle height generally has minor impact on the investigated response variables.